I automate business processes and build AI assistants and recommendation systems — from scratch, tailored to your business. I've delivered AI for 10+ companies: from clinic bookings to wholesale demand forecasting. I'll tell you honestly when AI adds value and when it's just expensive decoration.
I ask this of everyone who comes with the idea "we need AI". The answer is usually both, and that's fine. But it's important to understand the difference: there are cases where AI delivers a 35–45% work saving within six months, and there are cases where it becomes expensive decoration that sits unused.
In 2026 AI became affordable — automating a business process with native language support costs from €2,900, not €30,000 like a couple of years ago. But not every business needs it. On this page I lay out exact services, prices and real case studies — so you can decide whether AI fits your situation, or whether the money is better spent elsewhere.
If you get fewer than 50 customer queries a month, if support takes less than 2 hours a day, if your sales cycle is short and non-standard — automation will be expensive decoration. In that case it's better to invest in marketing or UX. If you're unsure about the technology itself, I'll say so plainly at the free consultation.
No hype, no "AI solves everything" mantra — only what I've already done for clients, built from scratch and tailored to your business.
I automate the concrete processes that eat time: document and email processing, lead qualification, reporting, content generation, translations.
A custom AI assistant trained on your data (RAG), with GPT-4o or Claude — answers 24/7, captures leads, integrates with your systems. Language-optimized, GDPR compliant.
Personalization for e-commerce or content platforms: "customers also bought", "you might like", cart recovery — all based on your real data. One case: +28% AOV.
Need a new mobile/web app with AI features from scratch? That's covered by app development — chatbots, recommendations, image and speech recognition in your app.
One-off build + a monthly API/hosting fee (€0–500/month depending on volume). Concrete proposal after a free consultation.
One process
Trained on your data
Multi-channel + integrations
I pick the final model after a POC, comparing them on your real data. Here's how they differ:
| Criterion | GPT-4o (OpenAI) | Claude 3.5 Sonnet (Anthropic) | Gemini Pro (Google) |
|---|---|---|---|
| Price per query | ~€0.005 | ~€0.008 | ~€0.003 (cheapest) |
| European languages | Great | Great, fewer errors | Good |
| Accuracy | High | Highest, fewer hallucinations | High |
| Context window | 128K tokens | 200K tokens | 1M tokens |
| Best for | General use, largest ecosystem | Accuracy, complex logic, long context | Cheaper tasks, large texts |
| EU hosting | Via OpenAI EU | AWS Frankfurt (Bedrock) | Google EU regions |
In practice I use Claude 3.5 Sonnet on 60% of projects (fewer hallucinations), GPT-4o on 30% (ecosystem integrations), Gemini Pro on 10% (cheaper at large volume).
Anonymized examples from my delivered projects — with concrete numbers.
An FAQ + booking assistant with native language optimization. Cut admin time by 45% — patients book themselves through chat, integrated with the clinic calendar.
Custom recommendation system + personalized cart-recovery emails with AI-generated copy and product selection. Tripled cart recovery in 3 months.
Recommendation system in the mobile app: "customers also ordered", "add to your order", personalized banners based on user history. Average order value grew over 4 months.
AI demand forecasting. Write-offs down 35%, out-of-stock situations down 60%. ROI in 7 months. Inventory managed by season, weekday and promotion data.
A GPT-4o assistant on the website with insurance-specific context. Reduced call-center load (answers simple questions 24/7) while website leads grew — instant answers convert.
Internal AI assistant for staff with RAG (company docs, processes, templates). New hires ask and get instant answers from the knowledge base. Onboarding 4× faster.
I never start building without a POC. It saves you thousands if the model doesn't fit.
We write down the concrete use case and ROI assumptions, then I test 2–3 models (GPT-4o, Claude, Gemini) on your real data. I'll tell you if AI doesn't pay off in your case.
Full implementation on the chosen architecture: RAG on your data, integrations, UI. You see progress in real time.
Deploy, documentation, team training. Then — analytics and prompt tuning based on real usage.
"Built an e-shop and deployed a chatbot. Knows his craft."
"Ordered a website for my company, done fast and well. I'm planning to add a chatbot next, so I'll be back — I liked everything."
"Very professionally done, everything on schedule, communication 10/10, fast and smooth. The functionality was even better than expected. Thank you."
Real client reviews — check paslaugos.lt →
Automating a single process — from €2,900. A custom AI assistant with GPT-4o or Claude, RAG and integration — from €4,900. A full AI solution with multi-channel and deep integrations — from €9,900. On top of the project cost there is a monthly API/hosting fee (€0–500/month depending on volume). The final price is set after a free consultation.
Yes. GPT-4o, Claude 3.5 Sonnet and Gemini Pro understand and generate text natively in all major European languages — natural, not just understandable. Every solution is optimized for your language: terminology, tone, regional specifics.
Fully replace — no. But AI handles 60–80% of repetitive queries (pricing, availability, bookings, FAQ) faster and 24/7. Staff then focus on complex cases. One case: AI cut calls by 35% and increased website leads by 28%.
Depends on the task. GPT-4o (~€0.005/query) — universal, largest ecosystem. Claude 3.5 Sonnet (~€0.008) — more accurate, fewer hallucinations, better for long context. Gemini Pro (~€0.003) — cheapest, good for simple tasks. I pick the final model after a POC on your real data.
Automating one process — 2–3 weeks. An AI assistant with RAG and context — 4–6 weeks. A full AI solution with multi-channel and CRM integration — 8–12 weeks. The exact timeline is set after the initial consultation.
Yes. I build a new AI layer that runs alongside your existing system and uses your data via API — no rewrite needed. For example, I add AI recommendations to an e-shop, automatic classification to a document system, or AI reply suggestions to a CRM. Deployment 2–4 weeks.
Typically 6–12 months, depending on the use case. Real examples: a clinic — AI paid back in 4 months (−45% admin time). A food wholesaler — AI demand forecasting in 7 months (~€45,000 annual savings). An insurance company — 5 months (−35% calls, +28% leads).
Yes. Every AI solution I build is GDPR compliant: data stored in EU regions, used only for the specific purpose (data minimization), users informed about AI use, consents logged. If needed, I use EU-hosted models only (Claude via AWS Frankfurt or self-hosted open-source variants).
iOS, Android, PWA from €2,900 — with AI integrations.
E-commerceBespoke custom commerce with AI recommendations.
WebsitesCustom Next.js sites with AI chatbot integration.
CRMCustom CRM with an AI agent in your sales pipeline.
Free consultation: I'll tell you honestly whether AI fits your situation, which solution is best, and what ROI is realistic. No commitment, no sales pressure.